The Big Sort

by Bill Bishop

  Was there something of the time that caused political segregation in the first Washington community? The cultural differences between nineteenth-century Bostonians and Kentucky frontiersmen were likely greater than the social gap today between San Franciscans and southern West Virginians. Or are there psychological effects common to like-minded groups regardless of the century? These aren't questions just for historians. They are real concerns today, and not only because we live in a country where nearly half the voters live in communities where presidential elections are preordained. Two hundred years after Washington, D.C., first emerged from the coastal marshes, the ideological boarding houses are back.

  In 1990, a young conservative Republican from suburban Pittsburgh named Rick Santorum challenged a seven-term Democratic incumbent, Doug Walgren. Santorum wasn't given much of a chance to defeat Walgren. In an early October poll, Walgren led 41 to 25 percent. Over the last few weeks of the campaign, however, Santorum ran a simple television ad. It showed a large white house—Walgren's house. "There's something strange about this house," an announcer said. The house was "strange" because it was located in McLean, Virginia, "the wealthiest area of Virginia," and not suburban Pittsburgh. "Maybe that's why he voted for a pay raise seven times," Santorum's ad observed. Walgren countered that he had bought a house in Virginia to "keep the family together," but the damage had been done.9 When Santorum unseated the Democrat, the social life of Washington, D.C., changed.*

  Congress is keenly attuned to survival, and members learned a lesson from Santorum's victory. "Now you don't move your family to Washington," former Republican representative Vin Weber said in late 2004. "Now you live in sort of a dormitory with members of your own party." Former Democratic congressman David Skaggs said that Newt Gingrich advised his Republican colleagues in 1994 not to appear settled in the District, leading to the rise of what Skaggs described as "dormitory life" in the capital.10 For example, California representative George Miller shares his house on Capitol Hill with New York senator Charles Schumer, Massachusetts representative Bill Delahunt, and Illinois senator Richard Durbin.11 In 2006, the combined household of Democrats scored higher than 90 percent on a voting ledger kept by the liberal Americans for Democratic Action, a record more in lockstep than that of the messmates of the early 1800s. After midterm losses in 2006, the homes of former Republican House members went up for sale at 129, 131, 132, 135, and 137 D Street SE.12

  The new boarding house norms discourage social interaction among legislators, intensifying the city's growing ideological isolation. These days, "the interactions that occurred over many decades between members, after hours ... and on weekends and with their spouses, simply does not occur anymore," said Weber.13 Fifty years ago, House Speaker Sam Rayburn, a Texas Democrat, served drinks at the end of the day to his Republican adversaries. Later, Republican leader Robert Michel, Democrat Dan Rostenkowski, and Republican Harold Collier shared a car on the long drive between Illinois and Washington, D.C.14 Disputes were mediated by a culture that encouraged a kind of heterogeneous civility. Former University of Chicago Divinity School dean Martin Marty was an intern at a church in Washington in the late 1940s when a parishioner, a congressman from Illinois, scuffled with a congressman from Mississippi on the floor of the House. "The following Sunday, their two pastors had the two families together at some club," Marty told me. "Why? Because Monday they had to be talking to each other again. Well, now that doesn't happen."

  Today, Weber noted, "they're on the last plane in before the first vote is cast on Tuesday, and the first plane out after the last vote on Thursday, and in between they sort of live in a dormitory, with people usually of the same party." Just as legislators brought the social segregation of their geographically isolated communities to the capital in 1800, many members of Congress today reflect the political segregation of American communities. "There is an analogue to road rage in the demeanor, the predispositions of newly elected members of Congress who are coming out of a society that is more self-isolating, more self-absorbed, less moderate in their day-to-day relationships," former congressman Skaggs said.15

  Social psychologists began studying the effects of groups on individuals more than one hundred years ago. That people living in homogeneous groups would be loath to compromise—or would even exhibit a bit of ideological "road rage"—would not surprise these researchers in the slightest.

  Judging Johnny Rocco

  It was Norman Triplett's enthusiasm for bicycle racing that, in 1897, led to the first experiment in social psychology. The Indiana University professor studied times compiled by the Racing Board of the League of American Wheelman for "over 2,000 racing wheelmen, all ambitious to make records." He found that riders racing against others posted faster times than when they pedaled only against the clock. Triplett gave all sorts of possible reasons why a group might have this effect on individual riders—suction, "brain worry," hypnotic suggestion—before settling on the Jules Verne—sounding theory of "dynamogenics." He proposed that the "bodily presence of another rider is a stimulus to the racer in arousing the competitive instinct." The presence of others on the racecourse freed up "nervous energy" that a solitary wheelman could not muster; the group served as an "inspiration to greater effort." To test his theory, Triplett rigged an experiment with fishing reels to determine whether children retrieved silk thread faster alone or in competition. As with the bicyclists, the ten- and eleven-year-old children worked the reels faster in the presence of others.16

  Triplett showed that groups change individuals in ways that individuals don't change on their own. And his results set off a century of experiments into the effects of group on individual, with results both surprising and disturbing. These studies found that people's opinions were deeply affected by groups, by notions of prestige, and by the opinions of majorities.

  In an early homemade experiment, a German schoolteacher first determined who the popular and unpopular children in a school gym class were. The teacher secretly instructed the popular group to disobey clear instructions during class. When the entire class was asked to raise their right hands, the popular children disobeyed and raised their left. After the class, however, the children in the class reported that it had been the unpopular children who had not followed instructions.17

  Muzafer Sherif asked Harvard and Radcliffe students to rank a list of sixteen well-known authors, including Joseph Conrad, Charles Dickens, Leo Tolstoy, and Mark Twain. The psychologist then gave the students sixteen paragraphs with the name of one of the famous authors attached to each. Sherif told them that the various authors had written the paragraphs, but in reality Robert Louis Stevenson had written them all. He asked the students to rank the paragraphs according to literary worth. The students ranked them in nearly the same way they had earlier judged the sixteen authors. When Sherif conducted the same experiment in Turkey, the Turkish students did the same thing.

  The experiments conducted by the German teacher and Sherif disclosed quite a bit about the fragility of human integrity. They also showed something important in understanding politics: what we think of what we hear or see or read depends largely on who said it, did it, or wrote it, and we are likely to find evidence that confirms our preconceptions.18

  In a 1951 study,19 Stanley Schachter divided students enrolled in economics classes at the University of Michigan into four groups, or clubs. There was a radio club, a theater club, a movie club, and so on. Each group had three undercover members planted by the research team. One of the undercover researchers read the story of Johnny Rocco, a juvenile delinquent awaiting a court sentence on a minor crime. (Yes, Edward G. Robinson played Johnny Rocco in the 1948 movie Key Largo.) The reader in each group asked the members what should be done with young Johnny. The clubs were required to apply a seven-point scale of punishment, ranging from something akin to hugs to hanging. The group members talked about Johnny for forty-five minutes. One of the other two undercover researchers was the "slider." The slider began the discussion at one extr
eme and gradually moved to what she perceived to be the middle of the group. The slider appeared to be convinced by the group's thinking. The third undercover researcher played the "deviant." The deviant determined which way the group was leaning and then took the position at the other extreme, maintaining that opposing view throughout the discussion.

  At the end of the debate, the club members were asked to make two other decisions in addition to punishment. First, they were to nominate members for an "executive committee" of clubs, clearly a position of honor. Second, they were told that the size of their club might need to be reduced. To help weed out members, they were asked to rank their preference for who should remain in the group. The deviant didn't fare well in these decisions. He wasn't picked for the executive committee and was consistently ranked low on the list of who should remain in the group. (Meanwhile, the slider was fully accepted by the group.) In all of the groups, however, rejection began long before any lists were made. As the deviant revealed himself in the discussion, group members gradually excluded him from the conversation. Eventually, they stopped talking to him altogether, effectively turning him into a nonperson. Schachter devised the test so that two of the four groups were made up of like-minded people. (Students who had a strong interest in movies or radio were placed in the same groups.) All of the groups excluded the deviant, but the more homogeneous groups were more intent on excluding the deviant than were the groups made up of a mix of students. The like-minded groups were quicker to stop talking to the person with the contrary opinion and rated him lower on the preference list for club membership.*

  A "Risky Shift" to the Extreme

  There was nothing sinister in the reactions these early experiments uncovered. People were responding to an innate need: to find safety in groups. "From our earliest moments on earth, we come to associate a wide array of positive outcomes with acceptance and love from others," psychologist Robert Baron, a professor at the University of Iowa, told me. "Right from day one, you form this very generalized belief that it is always bad to disagree with others." Beginning in the 1960s, however, social psychologists came to understand that like-minded groups not only enforced conformity but also tended to grow more extreme.

  The discovery began with a misdirection. In 1961, a graduate student named James Stoner asked subjects in an experiment to consider the prospects of George, a competent chess player who has the misfortune of drawing a top-ranked player in a tournament's early round. The game begins, and George sees an opportunity to attempt a risky play that could bring quick victory. If it failed, however, it would result in certain defeat. The subjects were then asked if George should attempt the risky play if there was a 10 percent chance of success, a 20 percent chance, and so on. The subjects decided individually at what odds George should try the maneuver. They were then asked to discuss as a group what George should do and arrive at a joint decision. What Stoner found—and what other researchers around the world would also find in subsequent experiments—was that the group always made a riskier recommendation than the average of the individual decisions. If, for example, the average of the individual judgments was that George should try the play if there was a 30 percent chance of success, the group would agree that George should take the risk if it paid off only 20 percent of the time.

  The consistent finding in this experiment became known as the "risky shift phenomenon," and as a piece of social psychological research, it was both provocative and deceptive. When the experiment was repeated in different ways and in different countries, researchers noticed a kink in the risky shift. In the chess game situation, most people thought the overmatched George should take a risk. But what if the hypothetical game was played from the point of view of the chess champion? In this scenario, individuals in the group leaned toward a restrained approach, and the group decision was more conservative than the average of the individual answers. Although the group decision still shifted from the average, in this case it became more risk averse.

  Social psychologists concluded by the end of the 1960s that what Stoner had discovered in his chess tournament experiment was the phenomenon of group polarization—that groups over time become more extreme in the direction of the average opinion of individual group members. Stoner's chess tournament advisers were inclined as individuals toward a risky play, and so their group decision was even riskier. In a different setting, where individuals were cautious, the group arrived at an even more cautious decision. Either way, the effect of discussion was to push the group and the individuals toward the extreme.20

  In another experiment, students in their last year at a Parisian lycée were asked their feelings about the United States and General Charles de Gaulle. After discussion, the students' positive feelings toward the general increased, as did their less than favorable inclinations toward the United States. The group didn't settle on the average of what the students thought as individuals. Instead, it adopted a more extreme position. Conventional wisdom is that group discussion balances out different points of view, but these researchers found that "society not only moderates ideas [but] it radicalizes them as well."21

  There have been hundreds of group polarization experiments, all finding that like-minded groups, over time, grow more extreme in the direction of the majority view. In one experiment, freshmen who joined fraternities were more conservative than freshmen who didn't. Senior fraternity members, however, were more conservative than freshmen. Freshmen who didn't join fraternities were more liberal, and the ideological gap between them and fraternity members widened during their years in college. In another experiment, people who were racially prejudiced became more prejudiced as they talked about race relations. In a third, intervention programs that clumped delinquents with other delinquents increased the group rate of law breaking.22

  Even people who are impartial by training are subject to group polarization. University of Chicago law professor Cass Sunstein and University of California, San Diego, business professor David Schkade reviewed decisions of federal court of appeals panels. The panels consisted of three judges, all appointed by either Republican or Democratic presidents. Sunstein and Schkade used the difference in political sponsorship to test whether ideology mattered in the panels'decisions. It did. All-Republican panels were far more likely to side with companies in labor or environmental cases. All-Democratic panels were far more likely to find against companies in environmental, labor, and sex discrimination cases.* Perhaps it's not surprising that Republican-appointed judges have views different from Democratic-appointed judges. But Sunstein and Schkade found that the same judges would shift their positions depending on the ideological makeup of the panel. A Republican-appointed judge sitting with two other Republican appointees voted more conservatively than when the same judge sat with a mix of Democrats and Republicans. A Democratic appointee would shift to the right when sitting with Republican appointees and would vote far more liberally when sitting with two Democratic appointees.23

  The lesson for politics and culture is pretty clear: It doesn't seem to matter if you're a frat boy, a French high school student, a petty criminal, or a federal appeals court judge. Mixed company moderates; like-minded company polarizes. Heterogeneous communities restrain group excesses; homogeneous communities march toward the extremes.†

  Social psychologists have proposed several theories to explain group polarization. Two have survived scientific scrutiny. The first holds that people in single-minded groups are privy to a large pool of ideas and arguments supporting the dominant position of the group. If there are good arguments in favor of the group's inclination, everyone hears them, and hears them often. Moreover, as the group talks about these ideas and arguments, individuals feel more strongly about them. People are more committed to a position once they voice it. The second theory holds that people are constantly comparing their beliefs and actions to those of the group. When a person learns that others in the group share his or her general beliefs, he or she finds it socially advantageous to adopt a position
slightly more extreme than the group average. It's a safe way to stand out from the crowd. It brings notice and even approbation.24

  "It's an image maintenance kind of thing," explained social psychologist Robert Baron. Everyone wants to be a member in good standing with the dominant group position. It's counterintuitive, but people grow more extreme within homogeneous groups as a way to conform. "One way to make sure you aren't mistaken for one of those 'other people' is to be slightly ahead of the pack in terms of your Republican-ness," Baron said. "It's hard to be a moderate Republican or a moderate Democrat, in other words, because you're afraid that other people will call you whatever. In racial terms, you'd be called an Oreo if you [were] black [but went along with whites]."25 Saint Paul knew this. Before his conversion, Paul said, "Beyond measure I persecuted the church of God, and wasted it." The disciple knew that going overboard in his pursuit of Christians served him well in the homogeneous society of the Jews. Paul explained that he "profited in the Jews' religion above many my equals in mine own nation, being more exceedingly zealous of the traditions of my fathers."26 Or, as Holly Golightly put it in the movie Breakfast at Tiffany's, "It's useful being top banana in the shock department."

  Like-minded groups create a kind of self-propelled, self-reinforcing loop. Group members send signals bolstering existing beliefs as they all vie to stand out as the most Republican or most Democratic in the group. And that sets off a new round of unspoken competition. Any successful talk radio host has realized, like Paul, that acclaim (and ratings) accrue to the most zealous. It's not enough to disagree with Bill Clinton or George W. Bush and to work for his defeat. These days, you must call for him to be impeached.